ChatGPT's free users can now also use GPT-5.6 Luna, with unlimited usage.



ChatGPT's default model has been changed from '

GPT-5.5 Instant ' to ' GPT-5.6 Luna '. This is said to reduce errors compared to GPT-5.5 Instant.

Improving GPT‑5.6 Sol in ChatGPT—and expanding access to GPT-5.6 Luna for free users | OpenAI
https://openai.com/index/improving-gpt-5-6-sol-in-chatgpt/




For free ChatGPT users, GPT-5.5 Instant was previously the default model, and conversations with the GPT-5.6 series were a feature reserved for paid plans. Now, GPT-5.6 Luna has been adopted as the default model, allowing free users to converse with GPT-5.6 Luna without limits.

OpenAI has reported, based on its internal evaluation of the performance differences between GPT-5.6 Luna and GPT-5.5 Instant, that 'GPT-5.6 Luna had approximately 62% fewer responses containing factual errors related to finance, medicine, and law compared to GPT-5.5 Instant.'

Additionally, a new 'Think' button has been unlocked for free users. Pressing this button on questions that require deeper thinking will cause GPT-5.6 Luna to think for an even longer period of time.



Plus and Pro users, who subscribe to the paid version of ChatGPT, can now use a newly added slider in the web, mobile, and desktop versions of ChatGPT to choose how much time ChatGPT spends considering their answers. It's recommended to set the slider short for everyday questions and long for planning, researching, writing, coding, or decision-making that requires more deliberation.



In addition, the 'GPT-5.6 Sol' available to paid users has also been updated to provide more direct answers. For simple questions, it provides direct answers including the necessary background information, while for more complex tasks such as multi-stage planning, research, and writing, it provides more detailed answers with clearly defined key recommendations.

GPT-5.6 Sol is designed to reduce errors by making better use of the information sources found when answers depend on dates, numbers, sources, rules, assumptions, etc.

OpenAI stated, 'This is a concrete step towards achieving richer intelligence. It will make the latest models available to more people, improve the usefulness and reliability of responses, and allow free users to continue text chat without rate restrictions.'

Related Posts:

in AI, Posted by log1p_kr